Acoustic or Elastic?

More Complete Is Not Always Right

The elastic engine is the more powerful of the two. It carries shear waves, mode conversion, and the full angle-dependent Zoeppritz amplitude that AVO reads, none of which the acoustic engine has. It is tempting to conclude that you should always use it. That conclusion is wrong, and correcting it is the whole point of this course.

The elastic engine costs more. It runs several times heavier, it needs shear velocities you may not have, and it introduces parameters that can go wrong. If your question never touches the extra physics, all of that is wasted. The right engine is the simplest one that still carries the physics your answer depends on.

The Decision, Task by Task

Work through the tasks and a clean line appears. A migrated structural image, or a fault-detection training set for machine learning, needs only geometry and P reflectivity: the acoustic engine, or even plain convolution, is fit for purpose, and elastic modelling would just burn compute. A direct hydrocarbon indicator, a far-offset AVO study, a time-lapse fluid change, or a converted-wave feasibility test all live in the elastic, angle-dependent response: here the acoustic engine does not merely cost less, it gives a confident and wrong answer, because it has no way to represent the fluid or the offset behaviour.

That asymmetry is the lesson. Using an over-powered engine wastes money; using an under-powered one produces a plausible result that is quietly false, which is far more dangerous. Fit-for-purpose modelling is the discipline of knowing which physics your question depends on and choosing exactly that. Part 7 opens a new axis the acoustic and isotropic-elastic engines both ignore, the dependence of velocity on the direction of travel: anisotropy.